Re: address pair exploration, flooding and state loss

[Erik Nordmark <erik.nordmark@sun.com>]

marcelo bagnulo braun wrote:

> so the conclusion is that reboots and lost state have to be detected at 
> least as fast as what is available today

yep

> may be but it does constraint the possible heuristics used for 
> establishing shim sessions.
> Maybe it would be enough to mention that the heuristics for establishing 
> shim sessions also are used for recover from those situations where the 
> context has been lost. However, i guess that need to take into account 
> where the heuristics for establsihing shim sessions are not useful for 
> recovering from lost state and provide a worst case recovery for this case.

Yes, it makes sense to describe this part of state recovery, and any 
implications it has on the recommendations for the heuristics used to 
determine when to establish shim state.

> so, so far what we have is:
>
> - before a rehoming event, packets may not be identified as belonging to 
> a shim session. If this is the case, the data packets associated to this 
> sessions are not useful to detect context loss, so alternative 
> mechanims, like using the heuristics for establishing shim sessions are 
> used for recover from this situations
> - after a rehoming event, context loss is detected upon the reception of 
> any packet associated with the shim session, whether signalling packet 
> or data packet. for that, data packets need to carry at least one bit 
> that identifies them as belonging to a shim session.

Makes sense to me.

   Erik